Calibration and Testing

Abstract

Seismic instrument will not always work perfectly and even new instruments might have faults. It is therefore important to be able to test the instruments. This includes checking and/or determining the response of the instruments, since knowledge of the correct response is essential for analyzing the data. Seismic equipment can be complicated so not everything can be checked, however simple tests will often be sufficient in most cases. The instruments needed for most tests are one or several of the following: a multimeter, an oscilloscope, a signal generator and a recorder.

Sensors: For passive sensors like short period seismometers, the following parameters can be determined: Free period, damping, generator constant and coil resistance. With these parameters the characteristics of the sensor is completely known and the response function can be calculated. For active sensors there will be no coil resistance. Many sensors also have a calibration input so the response function can be measured directly inputting a signal, either from a signal generator or in the form of a calibration pulse and recording the output with a recorder. An alternative method to get the generator constant and the response function is to use a shaking table. In that case the ground input is absolutely known and the response function is directly related to the output. However, shaking tables are not common. An alternative is to use the shaking of the ground as input and measure the output from a known and an unknown sensor and compare. A force balanced accelerometer can be checked for static parameters using the gravitational field and just incline the sensor.

Digitizer and/or recorder: The frequency response can be measured with a signal generator as input and just recording the signal. The sensitivity will then also be calculated from the known input and the maximum input is found by increasing the input voltage until clipping occurs.

Noise in instruments: The instruments always generate noise and ideally it should be smaller than the signal noise. This is often the case for the best digitizers but sensors will always have noise that limits its usefulness, particularly at low frequencies. Measuring the self noise of an instrument is therefore an important task. For digitizers, the simplest test is to short the input and measure how much noise is recorded. For sensors, a similar test might be to lock the sensor, if possible, and then observe the noise. A very popular approach is the coherency method. Under the assumption that two sensors or digitizers have identical response, the output from the two instruments recording the same input signal will differ only in the instrumental self noise. One then determines the coherence between the two records and assumes that the coherent signal is the seismic signal and the incoherent signal is instrumental noise. With three instruments, the response does not have to be known.

Several practical examples of testing and calibration are given.